Almost every organisation I walk into right now is trying to change faster. New tools, new workflows, new expectations, landing weekly. And almost everywhere, leaders hit the same wall, not the technology, but the people who are meant to use it.
The usual explanation is culture. People are hesitant, sceptical, slow to adopt; fix the mindset and the change will follow.
I think that gets it backwards.
When someone resists change, they’re usually not being difficult. Their brain is doing exactly what it evolved to do, protect them from uncertainty. That isn’t a culture problem. It’s biology, and it’s the thing most change strategies quietly step around.
I spent an hour on this in a recent Singularity University discussion. The short version: even people who genuinely want the future can feel threatened by it. And in the AI era, that gap is about to matter more than it ever has.
Change is a prediction error
Your brain runs on prediction. It’s a forecasting machine, quietly guessing what comes next so it doesn’t have to work everything out from scratch. When the world matches the forecast, you get a small hit of dopamine and a comfortable sense that things are under control. When it doesn’t, you get the opposite, a drop.
A small drop is useful. It’s how we learn: something surprised me, pay attention. The trouble starts when the drops stop being small.
Think about playing table tennis as a kid against someone older and better. You lose the point. Then another. Then another. Pretty quickly you stop trying to win and just want it to be over. That’s a brain handed too many prediction errors, too fast, and it’s exactly what a badly run change program does to a workforce. Overwhelm, then helplessness, then the quiet conclusion that nothing you do will make a difference. Cortisol, not curiosity.
The leader’s job isn’t to remove the drop. Change without discomfort isn’t change. The job is to calibrate it.
Urgency is not the lever you think it is
Here’s what I keep seeing. Leaders feel the pressure of AI, the market’s moving, competitors are experimenting, so they reach for urgency. Everything is changing. We have to move faster. This is existential.
They think they’re creating momentum. Often they’re just turning up the threat dial on every nervous system in the building.
And here’s the part that surprises people: the most engaged organisations are frequently the most fragile. Everyone cares, everyone’s straining to keep up, everyone’s waking up already stressed about the pace, running on cortisol. That looks like engagement. It isn’t adaptability. Gartner found willingness to support change fell from 74% in 2016 to 43% by 2022, while the number of major changes landing on the average employee went from two a year to ten. We didn’t get lazier. We got flooded.
The fix is almost embarrassingly simple, and almost nobody does it. For every urgency signal you send, send a stability signal to match. Here’s what’s changing. Here’s what isn’t.
Your people don’t need you to be certain, you can’t be, and they know it. They need you to be predictable. It’s the same move you make at home: honey, I love you, but I’m working late Friday, I’ve got a deadline. The reassurance is what lets the disruption land. A nervous system that knows where the solid ground is will take a risk. One that doesn’t, won’t.
AI doesn’t threaten your job. It threatens your competence.
Most technology rollouts changed where you did the work. AI changes what being good at the work even means.
That’s a different kind of fear. For a lot of people it isn’t “AI will take my job.” It’s quieter and more corrosive: I don’t know how to be good at my job anymore. Train someone on prompts and tools and you’ve addressed none of that.
This is where the real divide of the AI era is opening, not between people and machines, but between two kinds of people using the same machines. Some are learning to direct these tools: prompting, arguing with the output, throwing away the first answer, applying judgment. Others just consume whatever the tool hands back. Little gods and spectators, identical software, completely different futures.
You can’t move a frightened person from the spectator column into the director column. Fear narrows people, it’s the exact opposite of the open, experimental state that learning a genuinely new tool demands. So the emotional work isn’t soft. It’s the precondition for the adoption everyone says they want.
From change management to change as usual
The old change playbook assumed a destination. Announce it, explain it, train everyone, land the new normal. Done.
AI has no new normal. The tools keep moving; the destination keeps receding. So the question stops being “how do we get through this change?” and becomes “how do we build an organisation that can keep changing without burning its people to the ground?”
That’s what I mean by change as usual, not constant chaos, but adaptation built into the operating system. Stable core, plastic edge. People need to trust that some things hold, precisely so they can be brave about everything else.
There’s a sharp reason to get this right, and it isn’t sentimental. Push people to adopt under threat and they don’t only resist, some comply in the worst way, offloading their judgment onto the tool just to make the discomfort stop. We’ve seen this film before: hand everyone Google Maps and, a few years on, nobody can cross their own city without it. Do the same with thinking and you get a workforce that can operate the AI but has quietly stopped interrogating it. One 2025 study of 666 people found heavier reliance on AI tools tracked with weaker critical thinking, a strong negative correlation. The nervous system you rush is the one that stops thinking. That’s the opposite of what this era needs from people.
What leaders should actually do
AI is a technical problem. AI adoption is a human one, and the leaders who win it will be the ones who treat resistance as information rather than defiance.
When someone hesitates, don’t push harder. Ask what the hesitation is telling you. Is the goal unclear? Is the ground genuinely unstable? Are they missing the training, or just the time? Are they quietly afraid that the thing they were great at no longer counts? Hesitation is data. Most leaders throw it away.
None of this is about removing every unknown. You can’t, and pretending otherwise is its own kind of threat. The brain will always want clarity, agency and trust before it does its best work. That hasn’t shifted in a hundred thousand years, and AI won’t shift it now.
The organisations that thrive won’t be the ones that adopted first. They’ll be the ones that made it safe enough to keep thinking, keep learning and keep contributing while the ground moved. That’s the whole game.
